A short reader guide.

Nyalai examines the evidence behind systematic investment strategies, against five published criteria. Its first operational framework is an adversarial calibration and validation framework for quantitative finance. It renders binary verdicts (CALIBRATED or NOT-CALIBRATED) on trading strategies, ensemble portfolios, and predictive models submitted for validation. Verdicts are intended for public release, subject to the submitter's consent, which is collected after the verdict.

Nyalai's gate specifications and thresholds are published under Creative Commons Zero (CC0-1.0) on the Refusal Doctrine page, an editorial abridgment of the sealed text. Anyone may inspect, contest, adopt, or extend the methodology. What Nyalai sells is the discipline of applying the method rigorously and publishing the results, including its own failures.

CALIBRATED: the strategy passed all five gates of the Refusal Doctrine (statistical fragility, cross-regime robustness, guided-search safety, probabilistic calibration, bootstrap confidence). It is a statement that, as of the date of the verdict, the strategy has survived a documented adversarial process. It is not a guarantee of future returns.

NOT-CALIBRATED: the strategy failed at least one of the five gates. The verdict-form documents precisely which gate failed, the empirical evidence, and the diagnostic. A NOT-CALIBRATED verdict is a load-bearing outcome, not a rejection ; the diagnostic is the value the submitter receives.

Depends on gate complexity, data volume, and submitter responsiveness. The duration is defined case by case, according to scope and evidence requirements. The evaluation window is fixed at submission. Nyalai does not extend the observation window to accommodate a favorable result.

Two channels are open. Nyalai is in a pilot phase: one external strategy submission has been examined (Verdict-form 002), and no verdict has yet been commissioned by an allocator.

Strategy submitters (quantitative research teams, portfolio managers, systematic trading operations) submit their own strategy for adversarial validation. The verdict-form and its audit trail are published on nyalai.com/verdicts once the submitter has consented, after the verdict.

Institutional allocators can discuss an examination of a manager they are evaluating or currently allocating to; during the pilot phase, this path is open for discussion rather than ready to order. See /submit for both paths.

Nyalai operates on a per-verdict basis. Fee scales with submission complexity (data volume, model architecture, evaluation window, gate coverage). Fee is agreed in writing before validation begins and is independent of the verdict outcome ; a NOT-CALIBRATED verdict costs the same as a CALIBRATED one. During the pilot phase, scope, fees and timing are agreed in writing before work begins. Assessment criteria are published and are not negotiated per engagement. Detailed pricing on request via contact@nyalai.com.

Three reasons. First, a validation methodology that cannot be inspected cannot be trusted ; if Nyalai refused to explain the method, the verdicts would be indistinguishable from opinion. Second, the state of the industry benefits from open methodology ; if competitors adopt Nyalai's gates and improve them, allocators end up with better strategies deployed. Third, openness is a form of discipline ; publishing the method commits Nyalai to defending it in public, and any gate that is wrong will be caught by the readership.

Nyalai's commercial position does not depend on secrecy about the method. What Nyalai sells is the discipline of applying the method rigorously and publishing the results, including its own failures such as the historical evaluation of Synapse v0.3 (NOT-CALIBRATED).

Yes. Per Doctrine §11.4 Contestation, the submitter has thirty days from verdict publication to file a substantive objection. The objection is reviewed by a validator instance that did not contribute to the original verdict. The outcome is published as a public addendum to the original verdict-form.

External arbitration by a roster of qualified adversarial validators outside the Nyalai institution is planned as an escape valve and is not operational today (per §11.1 anti-consolidation, it awaits the emergence of at least one independent adversarial validator institution operating in parallel to Nyalai on comparable methodology).

Nyalai's archive of dated snapshots. archive.nyalai.com holds snapshots of the canonical pages (doctrine, verdict-forms, methodology), each with a MANIFEST.json of SHA256 hashes. It holds one snapshot today, labelled 2026-07-19. The snapshots of later versions have not been taken yet.

A silent secondary tier of off-domain preservation mechanisms is documented internally for post-Nyalai durability. The secondary tier is not surfaced from this page or from any other reader-facing Nyalai surface, to protect the reader experience from third-party chrome that off-domain preservation providers inject on their own pages.

Doctrine §6.6 (Non-manipulative Extension) requires a pre-blinding step that strips submitter identity, relational metadata, and prior-conflict fields from the package before it reaches the validator instance, and a scan step that searches preserved fields for embedded identity leaks. The strip list is a floor rather than a ceiling, and it is not published yet. Its application to a published verdict has not yet been documented: see the next paragraph.

The Doctrine specifies an Attestation panel documenting the protocol's completion : pre-blinded package SHA256 hash, exposure events (including the null case), identity of the second validator in case of recusal, and the timestamp of the pre-blinding step. The verdict-forms published to date do not carry this panel, and the full protocol is not published yet.

Institutional inquiries : contact@nyalai.com. Including one of the following subject-line prefixes helps identify the nature of an inquiry : governance question, verdict submission, allocator-led examination, external arbitration, methodology question, general inquiry.

For questions about a specific verdict-form, please reference the verdict identifier (for example, VF-002 or R-01) in the subject line.